Sensor networks at NRS reserves pioneer new ways to observe Earth

Throughout history, humans have searched for new ways to view the Earth and better understand its natural processes. Just as the earliest microscopes and telescopes extended our vision into previously unseen worlds, other technologies have had equally profound impacts. In the 1840s, the quest for a new perspective led some scientific explorers to send aloft still-cameras tethered to balloons in order to document an expanse of forest or city. Aerial photography from airplanes became common in the 1920s and '30s, capturing images of large swathes of the planet's surface. By the middle of the twentieth century, breakthroughs in space sciences raised to new levels our ability to view the Earth. The first weather satellites dramatically improved our ability to detect destructive storms and predict their courses. The 1972 launch of Landsat 1 ushered in a new era of multispectral observations that include visible light, near-infrared, mid-infrared, and thermal data. Today these satellites are tracking everything from typhoon destruction in Burma, to pollution plumes in Chesapeake Bay, to the retreat of Alaskan glaciers, to water-consumption patterns in the western United States.

As impressive as these accomplishments are, the laws of physics cannot be defied. As one's distance from an object increases, the ability to detect fine details diminishes. Remote sensing - identifying, observing, and measuring an object without coming into direct contact with it - has its limits. Engineers and environmental scientists have continued to search, with their feet on the ground, for ways to monitor natural processes from the Earth itself, without taking to outer space or even aerial space. The explosion of computer and networking technology in the 1990s has now put us within reach of this goal.

Today networking technology, pioneered by engineers and scientists working at NRS reserves, is playing a key role as we enter a new era of global discovery through monitoring. Networked -sensor technology prototyped, field-tested, and refined at the James San Jacinto Mountains Reserve in Riverside County has provided the basis for systems that monitor everything from local microclimates to continental climate change. Technology from the same reserve is now being assembled in Central America to support investigations into carbon flux in tropical rainforests. And sensors originally developed to monitor the elephant seal population at the Año Nuevo Island Reserve in San Mateo County have evolved to track the animals into the open ocean and reveal, for the first time, the rest of their life stories. These same sensors are also providing dramatic new insights into the functioning of the oceans themselves.

Within the NRS, sensor networks are proliferating. At Angelo Coast Range Reserve in Mendocino County, scientists from the Keck HydroWatch Project are using a wireless network to monitor a watershed's response to rainfall without leaving their offices in Berkeley. At Quail Ridge Reserve in Napa County, scientists plan to use the reserve's innovative wireless mesh network to monitor frog and bird populations. And at Sagehen Creek Field Station in Nevada County, a recently installed series of 13 towers will provide a broadband, wireless communications system for scientists working throughout the 8,000-acre watershed.

The CENS Story
At first glance, the newly established Blue Oak Ranch Reserve in the Mt. Hamilton range above San Jose would seem more notable for its natural beauty than its technological potential. Aside from the array of solar panels on the roof of the barn that serves as the reserve's temporary headquarters and a single weather and communications station on a nearby hilltop, technology appears to play a secondary role here.

But walk the hills with Reserve Director Mike Hamilton and he paints a very different portrait. Climbing the steep hillside to the weather station, he explains: "We call this the Alpha Node. It's a solar-powered weather station, but it's also a wireless relay point that links the Lick Observatory [owned and operated by UC and located on nearby Mt. Hamilton*] to a directional Wi-Fi radio that points down to the barn, providing us with Internet access. And this omni-directional antenna plugs into the router on the tower to create a large Wi-Fi cloud on the top of the hill that's strong enough to get a signal down to the pond and the stream at the foot of the hill, so researchers will be able to monitor these locations using portable wireless environmental sensing systems."

Obviously, Hamilton is comfortable speaking as both an ecologist and an engineer. From 2002 through 2007, he served as a principal investigator for the Center for Embedded Network Sensing (CENS), a National Science Foundation-funded Science and Technology Center based at UCLA and involving faculty from university campuses throughout Southern California. At the time, Hamilton was director of an NRS reserve administered through UC Riverside, the James San Jacinto Mountains Reserve, where many CENS concepts were field-tested.

Now in its seventh year, the CENS Project has had a major influence on ecological observatory networks throughout the world. "It's such a huge field of integration of interdisciplinary science between engineers and computer scientists and environmental scientists," notes Hamilton. "It seems that everyone is doing sensor networks today, and a lot of what they are doing is based on the work we did at the James Reserve. We've essentially provided the specifications for how to do things. Deborah Estrin (Director of CENS) has always had a very open-source approach to the computer code, and the details of how to build our systems are all published either in our own reports or in other publications. Over a thousand publications came out of CENS - and the James Reserve has a credit in many of them. We have an astounding publications record in that area of science."

Concepts pioneered by CENS have influenced the ongoing development of the National Ecological Observatory Network (NEON), a continental-scale research platform for discovering and understanding the impacts of climate change, land-use change, and the encroachment of invasive species. Tentatively scheduled to begin operation in 2013, NEON will include 20 fully instrumented ecological observatories across the country, as well as mobile instrument packages that can be moved to areas of special interest. All of these data sources will be linked by an advanced cyber-infrastructure designed to record and store data for the next 30 years. Hamilton and Estrin both served on the NEON technology committee that helped to establish the types of technology that would be used at each NEON locale.

"We were involved in the original prototyping design of what NEON should become," Hamilton explains. "Our committee solicited input from hundreds of ecologists, the potential end-users for NEON. We realized that all major NEON sites would have sensor networks associated with them, as well as large instrumented towers that would measure CO2 fluctuations very precisely to provide a sense of an ecosystem's productivity, how it breathes. Our job was just to develop the specifications. What did it need to do? What should its major components be? The model numbers, the brands of the systems they select, the installations are still being worked out by contract engineering firms."

In his new position at Blue Oak Ranch Reserve, Hamilton has the opportunity, and the challenge, of drawing on what he learned at the James Reserve to develop a network from scratch. His focus is on reliability and ease of use. He says: "I'm entering this era where I feel we should be getting more into ‘plug and play' on sensor networks, rather than research and development. Over the last seven years, CENS has worked out many of the engineering ideas. Now it's really about corporate partners taking those ideas and putting them into products. There's a lot of growth right now in using sensor systems for precision agriculture, ranging from viticulture to golf course irrigations. Those seem to be the big areas where embedded-sensing and mesh networks are playing out. Our field, ecological monitoring of microclimates across a diverse landscape, is a niche market. But the off-the-shelf systems need to be far more reliable and require less human effort to maintain than the prototype systems that we've been working on at the [NRS's] James, Stunt Ranch, and Quail Ridge reserves, where a small cadre of computer-science students keeps the systems running."

To encourage the development of commercial systems, Hamilton is drawing on Blue Oak Ranch Reserve's proximity to Silicon Valley to collaborate on new product development. Two major technology companies are currently installing test systems. "Both Sun Microsystems and Crossbow Technology are deploying small test beds here to show different applications for their tools," Hamilton explains. "Both groups are interested in relevant applications that they can then sell worldwide. So we're teaming them with faculty and students from UC Merced to monitor wetlands that support salamander populations by deploying sensor networks to measure the changes in rainfall, soil moisture, water depth, and some of the chemical parameters of the water, such as salinity, that vary across the reserve's ponds, depending on soil type and water source. They all have different populations of amphibians, and they're going to be different from pond to pond, so if we can set up these test beds in a few different wetlands, we can do comparisons across the reserve. We also want to test the reliability of the systems because, in the future, the reserves will want to pick those that prove their worth."

Moving in this same direction, CENS has worked with the Information Sciences Institute (ISI) at the University of Southern California to develop network products based on their research. SensorKits (), their first products, are standardized, flexible, networking systems that can be easily assembled and deployed in a wide variety of environments. CENS and ISI engineers have designed the units to be compatible with a wide variety of commercial sensors and to require a minimum of technical expertise to set up and maintain. Their modular design allows researchers to select the specific sensors they need for their project, connect to a SensorKit system, and literally carry the lightweight, waterproof units to their research locales. The kits are already being used by researchers at sites around the world, from Argentina to China.

Rundel in the Jungle
SensorKits and other CENS technology innovations field-tested at the James Reserve are key to a major sensor network development project now underway at the La Selva Biological Station in northeastern Costa Rica. Owned and operated by the Organization for Tropical Studies, a 63-member consortium of universities and research institutes, La Selva has become one of the world's most intensively studied rainforests.

As of February 2009, the project had been delayed a few months by logistical problems. Talking over the phone from his Los Angeles home, UCLA ecologist Philip Rundel sounded philosophical: "These things always take more time than you'd like. The towers are being built in Ireland. They have to be shipped across the ocean, make it through customs, and then somebody has to bring them to the field station and figure out how to set them up. The staff are already digging a 500-meter trench from the laboratory to the research site for the electricity and fiber-optic connections."

Rundel has been involved with CENS from its early days. "I focus on ecophysiology," he notes, "linking atmospheric processes to ecosystem processes. Sensor arrays are very relevant to our work. But remember, three-fourths of the CENS funding was for engineering. And terrestrial ecology is just one of the four application areas they're looking at. We've really been focused on the technology, and we're just now getting to the scientific questions."

When it's completed, La Selva Rainforest Ecological Portal will feature five interlinked towers with sensor arrays monitoring vertical and horizontal profiles of microclimates within the primary rainforest. Each tower provides a heavily instrumented array that will be directly connected, for both electrical power and data transmission, to the station's laboratory. Beyond the microclimate sensors there will be a series of pan-tilt-zoom cameras and acoustic sensors arrayed above and within the forest canopy. These cameras can be manipulated over the Internet to allow use by individual reviewers from a variety of research projects. All of the arrays will feed data to servers and hard drives in the station's climate-controlled Sensor Operations Laboratory, where data can flow directly to the Internet.

The architecture for the dense array will blanket a 4-hectare plot from the forest's floor to the top of its canopy. In addition to sensor nodes monitoring microclimates on the ground, researchers will be able to access the 30-meter-high canopy to collect samples and make observations via three stairway towers and a connecting canopy walkway. "One of the most challenging things in a rainforest is that there's more diversity off the ground than on the ground," Rundel notes. "The towers will give us the access we need to study that diversity."

Beyond providing access for researchers, the walk-up towers, along with two additional radio towers, will be linked with heavy wire cables. NIMS (Networked InfoMechanical Systems) mobile sensors, developed by CENS engineers, will move along these cables, raising and lowering sensors to take atmospheric samples at different elevations within the canopy.

Rundel, who studies ecosystem dynamics and carbon flux at La Selva, is confident that the arrays will provide invaluable new insights. "Though most people think of the tropical rainforests as sinks for capturing carbon emissions from the atmosphere," he explains, "some recent studies have suggested that this might not always be the case."

At La Selva, in particular, the trees aren't growing as fast as they are in the Amazon, and the rate of tree growth can be correlated with the amount of CO2 sequestered by a forest. Based on the station's long-term climate records, some researchers have suggested that this difference is due to the fact that current nighttime temperatures in the area are slightly warmer than they have been in the past. And if it's true that nighttime warming reduces a forest's ability to store carbon, then the Earth's warming pattern could accelerate.

A related question that interests Rundel is the role that gaps in the canopy, caused by fallen trees, might play in the amount of CO2 that reaches the atmosphere. In most cases, scientists monitor air turbulence above the canopy to determine how much CO2 a forest is emitting. But Rundel and other researchers suspect that the CO2 and other gases might gather under the canopy and then flow horizontally towards canopy gaps that act like chimneys. This is one of the theories researchers will be able to test once they can monitor the microclimates in these gaps.

The science is exciting, but so are the challenges. "Part of the challenge will be calibrating the equipment," Rundel notes. "Ants will get into everything, and it's extremely wet, with nearly 14 feet of annual rainfall. But the staff is excellent. CENS engineers have gone down there to train them, and the systems have proven reliable at the James Reserve through a lot of bad weather. It's going to be very interesting."

SealNet
One of the holy grails for terrestrial scientists has long been "smart" sensors that would detect and focus in on interesting ecological data. If an audio sensor heard a rare bird, for example, it would recognize the species and track the individual to find out more about its lifestyle. Marine scientist Dan Costa of UC Santa Cruz, chairman of the NRS Universitywide Advisory Committee, has already deployed an army of smart sensors that travel hundreds of miles to find areas of ecological interest. Once they find a spot, they'll often spend weeks there, providing a wealth of data on a previously unseen locale.

Costa's secret? Pelagic species that carry satellite tags. Say, for example, that an oceanographer is interested in how the waters under the Antarctic ice pack are responding to climate change. By equipping southern elephant seals (Mirounga leonina) with oceanographic sensors, researchers can measure ocean structure and water-mass changes during the elephant seals' months-long feeding forays to the Antarctic ice pack. Each time an animal surfaces, its tags transmit information to a passing satellite, providing almost real-time data on its location, its physical status, and the prevailing environmental conditions.

In association with biologist Barbara Block of Stanford University, Costa was a founding principal investigator for TOPP (Tagging of Pacific Predators, ). Costa has been a leader in an era of discovery that has, for the first time, revealed the secrets of a host of animals that spend their lives in and above the open ocean. "We knew so little about these animals," he explains, "our original hypothesis could only have been: does this animal live in the ocean? That was the level of our knowledge. For the elephant seals at Año Nuevo Island, we knew where they bred and reproduced, because that happened onshore. But that was it. When we put the earliest tags on them, we found that some of them went to Japan, because the tag would show up there. But what happened in between was a mystery."

In addition to revealing the behavioral and migratory patterns of dozens of pelagic species, the sensors have also begun to reveal basic patterns in the ocean itself. "There are usually two phases to animals' movements," explains Costa. "Either they're transitory, or they're stopping and feeding in places. When they're transitory, most biologists say there's not much to observe. But to a physical oceanographer, these data [collected from transitory animal movements] are the most interesting. Unlike the CENS program, which heavily instruments a location, we instrument individual animals and let them describe the environment in which they live."

"Physical oceanographic data are fundamental to understanding ocean weather," Costa continues. "The salinity profiles recorded by the sensors are how you measure weather in the ocean. Everybody understands temperature, high pressure, and low pressure in the air. In the ocean, temperature and salinity determinants are the equivalent. Our tags provide these data and so help to fill in some very large gaps in the larger global picture."

In addition to tracking oceanographic data, the tags also monitor the animals' physical characteristics. "They're telling us how hard these animals are working," notes Costa, "how far they swim, and how they respond. We already know that these animals are adapting to a changing environment. When it's warmer or colder, they're further from or closer to shore. That's the major part of my research: understanding how these animals are responding to the changing ocean. Are they working harder? How hard do they forage? Are the prey patches bigger or smaller? That gets back to baseline, and as we better define the habitat, we'll have a better framework to interpret what's going to happen to these animals as the environment changes. We're defining the physical features of the environment with what they do and where they live, so we're laying out a baseline for analysis."

Earlier this year, Costa achieved a career benchmark when he traveled to Isla Lobos in Uruguay to put tags on a population of South American sea lions (Otaria flavescens). "I've now worked with six of the seven sea lion species," he explains. "The seventh is the Japanese sea lion and it's extinct, so I don't have plans to tag or take blood samples from that species."

Much more than a personal accomplishment, however, Costa's experiences with the southern sea lion typify his approach to research. Currently, scientists know very little about the lives of the South American sea lions, except for what they could observe from places where the animals come ashore to breed and to bear young. "There have only been only a few papers on their behavior at sea," notes Costa. "People are unsure whether their population is stable or declining. So we're trying to find out where they're going, looking into their diving capabilities. The oceanographic data are really a secondary benefit when we begin looking at a new species. On the other hand, if you combined the data captured by all the sea lion and seal species, we now have oceanographic measurements that cover about half the Earth."

The success that Costa and dozens of colleagues have had with the TOPP program has exceeded even the most optimistic projections. As part of the global Census of Marine Life, TOPP is beginning to build a picture of the "other" two-thirds of the Earth's surface. "This wasn't random," says Costa. "We had questions we wanted to answer, and we specifically targeted the technology that would give us the data we need to answer those questions. But we still have a long way to go. It's one thing to collect all these data; it's another to make sense of them. The challenge now is to collate the data and combine them in ways that are visually accessible. [See illustration below.] That's our vision." Just as the James Reserve was a key test bed for the development of the CENS technologies, Año Nuevo Island Reserve has been key to the development of SealNet.

In contrast to their terrestrial colleagues, marine scientists have been forced to rely on sensor technology from the outset. They had no other options. "I think the sheer difficulty of doing what we do has led us to develop technology to a much greater degree than the terrestrial scientists," says Costa. "We're studying an animal that moves about the entire North Pacific, so there's a real difference in the scale at which we're asking questions. The funny thing about what I do is, I'm asking what happens between here and the Aleutian Islands, but I'm also asking questions at the scale of what happened every hour or at the last two meals. So we are getting both fine-scale and large-scale information, because that's also the scale over which the animals are living."

"We know so little about these animals that every opportunity to put tags on them to gather information about them is an incredible opportunity, and even more important in the context of their physical environment. So that's why I'm so excited about the idea of working with the oceanographers who want these data.' - JB

For more information...

...about CENS technology field-tested at James San Jacinto Mountains Reserve and now being installed at La Selva Biological Station in Costa Rica, contact:
Philip W. Rundel
Ecology and Evolutionary Biology
University of California, Los Angeles
Los Angeles, CA 90095
Phone: 310-825-4072
Email: rundel@biology.ucla.edu

...about pinniped tagging developed at Año Nuevo Island Reserve, contact:
Daniel P. Costa
Center for Ocean Health
100 Shaffer Road
University of California
Santa Cruz, CA 95060-5730
Phone: 831-459-2786
Email: costa@biology.ucsc.edu>

References
UC Natural Reserve System Special Research Projects: National centers & other landscape-scale projects that utilize NRS reserves. April 2008.
Page 6: Science & Technology Center (STC): Center for Embedded Networked Sensing (CENS).
Page 19: Tagging of Pacific Pelagics* (TOPP) [*Now changed to: "Tagging of Pacific Predators"].
Available as a downloadable PDF file at: .

M. P. Hamilton et al.
New Approaches in Embedded Networked Sensing for Terrestrial Ecological Laboratories. Environmental Engineering Science 24: 192-204 (2007).
M. F. Allen et al.
Soil Sensor Technology: Life within a Pixel. Bioscience 57: 859-867 (2007).
A. A. Shaffer et al.
Migratory Shearwaters Integrate Oceanic Resources across the Pacific Ocean in an Endless Summer. PNAS 103: 12799-12802 (2006).
J.-B. Charrassin et al. Southern Ocean Frontal Structure and Sea-ice Formation Rates Revealed by Elephant Seals. PNAS 105: 11634-11639.